Human-Centered Agenda for Discovery and Innovation
Statement to
President's Council of Advisors on Science and Technology
(PCAST)
Subcommittee on Networking and Information Technology
September 6, 2006 (revised version September 23, 2006)
Ben Shneiderman
Department of Computer Science
University of Maryland
College Park, MD 20742
Email: ben@cs.umd.edu Homepage : http://www.cs.umd.edu/~ben
Phone: (301) 405-2680 Fax: (301) 405-6707
Professor, Computer Science, Univ. of Maryland
Founding Director (1983-2000), Human-Computer Interaction Lab
Member, Institute for Advanced Computer Studies
http://www.cs.umd.edu/hcil
Introduction
Thank you for giving me the opportunity to comment on the NITRD program,
especially on the topic of Human-Computer Interaction and Information Management
(HCI&IM). While my training is broad within computer science, my career has been
happily filled with interdisciplinary projects that reach to many areas of science and
engineering, as well as to business, psychology, arts, and humanities.
My focus has been in developing the discipline of human-computer interaction and the
related applied field of usability engineering. In recent years my work has focused on
information visualization.
PCAST Statement revised
-
1 -
Ben Shneiderman 9/23/2006
The Next Generations of Information Technology
Computing professionals can be proud of the rapid evolution of computer science and
technology over the past 40 years: Moore’s Law tells the story of denser and more rapid
chips, personal computers spread these capabilities to many users, the Internet enabled
them to communicate, and Google provided remarkable capabilities to locate information.
These technologies will continue to mature and spread, but to envision the future, I ask
what the post-Moore’s Law world will be? What follows the personal computer? What
is the post-Internet Age? And what is better than Google?
These technologies might be best seen as the overture; and now the symphony begins.
The score of the symphony is being written by those innovative individuals who seek to
enhance human creativity, support collaboration within teams, and shape successful
communities. All these human-centered aspirations become possible because of well
designed user interfaces.
Good HCI makes life better (iPod, XboX, Wikipedia), accelerates business
(Amazon.com, ebay, Google) and prevents disasters (air traffic control, radiology); bad
HCI generates frustration (spam, pop-up ads, privacy violation) and causes disasters
(Airbus crashes, medical errors). HCI is increasingly recognized as the driving force in
many successful designs, and is taking a leadership role in national priorities such as
homeland security, health informatics, sustainable energy, and scientific research.
The maturity of the information, networking, communications, and computing field
means that the potent infrastructure is in place to enable compelling breakthroughs from
human-computer interaction research. A shift in national funding to accelerate
development of HCI could result in dramatic improvements in the quality of many
systems that enable discovery and innovation:
1) the rapid growth of computing power, usually described by Moore’s Law, has made
for profound changes, but its success means that other metrics become equally important.
For many applications, productivity and creativity are shaped by the quality of user
interfaces as well as by processor speed. A second metric of progress might be a Human
Capability Index that measures how rapidly an individual can accomplish benchmark
tasks, multiplied by the percent of individuals who have access to the required technology.
2) personal computers were a great success, but now users have higher expectations;
today’s technologies must be far more than personal, they need to be connected,
collaborative, and social. The expansion of open source software, Wikipedia, and social
tagging are shifting the balance from centralized producers to distributed innovation.
3) the Internet is valuable and remains important, but in the post-Internet world, the
impact of universally usable technologies such as cell phones, mobile devices, and
embedded systems is growing rapidly. Other high-impact technologies are media tools
for managing music, taking photos, and sharing videos. These emerged from consumer
PCAST Statement revised
-
2 -
Ben Shneiderman 9/23/2006
electronics but have powerful implications for information integration, metadata tagging
methods to create ontologies, and standardization to facilitate sharing. The power of
search engines and availability of exploratory search tools is changing the balance of who
has access to information.
4) Google’s search strategies are remarkable, but visualization tools will enable 100-fold
increases in the user’s capacity to explore information. Even in peta-sized genomic
databases, users will be able to discover relevant items, find surprising outliers, recognize
important clusters, and identify relevant relationships.
The shift to creativity support tools, distributed innovation teams, and social networking
communities is reshaping the experience and possibilities for every worker, voter, and
consumer. I’ve tried to capture this shift in the motto: The old computing is about what
computers could do; the new computing is about what people can do.
These changes are being recognized internationally, and influence policy discussions of
issues such as globalization, outsourcing, competitiveness, and international trade
balances. They also influence federal research and development projects, which now
have the potential to shift towards discovery and innovation. Many countries have
recognized these potentials (especially United Kingdom, European Union, China, Korea)
and are working through their scientific agencies and with appropriate corporations.
Shifting to new priorities
Accommodating these new realities requires a shift in priorities. This shift might be
reflected in small changes to the NITRD budget by rewording to replace older notions of
smart, cognitive, intelligent or autonomous systems, e.g. “systems that know what they
are doing. Cognitive systems able to learn”. These older notions are vague and do not
provide appropriate guidance for designers.
Once freed from these notions, the research directions could be aligned with futureoriented human-centered themes that support discovery and innovation in key
applications such as genomic data analysis, transportation safety, and medical informatics.
The key topics include creativity support tools for individuals, collaborative
environments for teams, social networking for communities. The guiding notions for
these power tools are comprehensible, consistent, predictable and controllable interfaces.
Then, the advantages will come from the huge numbers of users who can engage in and
benefit from sensemaking from intelligence data, cross-cultural communications for
diplomatic exchanges, and information sharing in product monitoring.
These goals will be achieved by understanding the current high level of frustration and
wasted time (38-45%) that have been measured in current systems. Reducing this drain
on national productivity is an important priority. Improvements to general purpose
applications will accelerate productivity gains for researchers and workers in every
PCAST Statement revised
-
3 -
Ben Shneiderman 9/23/2006
domain. Other improvements will bring advanced power tools for expert users based on
visual analytics, information visualization, and exploratory search.
The future key technologies will be personal medical devices, convenient mobile
devices, pervasive embedded devices, and widespread access to real-time remote sensors.
To make these transformations substantial investments are needed in security and privacy
technologies, as well as their implementation and maintenance. An important component
will be oversight strategies for individual users, corporate data aggregators, and
government agencies. Successful oversight will ensure high levels of performance, and
facilitate continuous improvement. Building trust, ensuring privacy, and clarifying
responsibility are the necessary lubricants for the next stage of expansion for networking
and information technologies, and the shift to discovery and innovation.
Three Projects promoting Discovery and Innovation
Three example projects may help to tell the story of how applications will be transformed
by creativity support tools for individuals, distributed innovation support for teams, and
social networking for communities:
Telescopes for High-Dimensional Data: Information visualization tools are becoming
available that enable rapid exploration of large databases for projects such as
pharmaceutical drug discovery, epidemic detection from emergency room admissions
data, and bio-surveillance from remote sensing devices. Dynamic queries for data
filtering is needed to provide 100msec updates for up to a billion records. These tools will
provide guided analytics based on rank-by-feature frameworks that use high-performance
interfaces to find patterns in bio-surveillance data, relationships in financial data, and
clusters in legal datasets. They will also enable discovery of outliers and anomalies in
fraud detection and terror prevention.
Community Response Grids: Community software can help prevent terrorism by
restoring social capital so participants are engaged, aware, and active. When Katrina-like
disasters or terror attacks occur, Community Response Grids facilitate rapid response,
integrated efforts, and endurance under stress. In the aftermath of disasters, Community
Response Grids accelerate recovery, facilitate cooperation, and restore normal services.
The resilience to disasters or terror attacks comes from well-established and trusting
relationships, so that when assistance and empathy are needed, they will be offered.
World Wide Med: If you are brought to a hospital emergency room anywhere in the
world, within 15 seconds your medical records will be onscreen in the local language.
Privacy and security must be given high priority to ensure adequate protection with
supervisory oversight. Lab tests or medical images captured locally will be uploaded for
inclusion in the patient record. A single button press establishes a communications link
with your physicians or their assistants. The World Wide Med becomes a tool to improve
patient care as well as a research platform for discovery of new medical knowledge.
PCAST Statement revised
-
4 -
Ben Shneiderman 9/23/2006
These three example projects could be expanded into dozens of projects in which humancentered thinking leads to new possibilities that are aligned with national priorities.
Research Agenda
The research agenda for discovery and innovation is large, but here is a starting list:
•
•
•
•
•
Mobile
devices
Embedded systems
Voting, Census surveys
Remote sensing
•
•
•
•
Exploratory search
Information visualization
Web-based info integration
Tagging, metadata collection
•
•
•
•
Privacy/security
Trust/responsibility
Universal usability
Online help
•
•
•
•
Social computing
Open source/open deliberation
Collaboration environments
Creativity support tools
These can be organized into categories relating to novel technologies that lead to
creativity support tools for individuals, collaboration environments for teams, and social
networking for communities.
Closing thoughts
Since HCI and Information Management are both vital areas, they would benefit to each
stand on their own. This would mean that each topic gets separate discussion that
clarifies the contributions of each technology.
To indicate the shift of attention, the phrase:
Networking and Information Technology Research and Development (NITRD)
Might be renamed as:
Discovery and Innovation Technology Research and Development (DITRD)
Another important step will be to increase the diversity in membership of PCAST and
NITAG, especially with regards to women and minorities. This is to ensure broadest
most thoughtful comments, and enable these scientists to gain the benefits of
participating.
A transformational step would be to increase education in HCI at the undergraduate and
graduate levels by increasing teacher training and fellowship opportunities. Each agency
PCAST Statement revised
-
5 -
Ben Shneiderman 9/23/2006
could also expand its educational offerings for current professionals so that humancentered thinking was part of every project.
Any management process is guided by appropriate measures of success. The shift from
technology metrics to human-centered metrics would be especially helpful. For example
in educational technology, progress is sometimes measured by the number of computers
per classroom or number of classrooms connected to the Internet. More appropriate
measures would be the number of web pages created by students, the number of emails
written in a foreign language, or the number of significant postings to wikis. Similarly, in
remote-sensing database projects, the number of petabytes is less important than the
number of users who make significant discoveries based on the data. For digital libraries,
the key is not how many electronic records are on the web, but how many users find them,
download them, and link to them.
HCI and human-centered thinking offer compelling approaches for directly responding to
societal challenges and addressing national priorities. Support for basic HCI research,
especially tied to the three example projects, would lead to high payoffs for these projects
and many related applications that depend on discovery and innovation.
Acknowledgements:
I appreciate the feedback and comments from these colleagues:
Ben Bederson, University of Maryland
Francois, Guimbretiere, University of Maryland
Harry Hochheiser, Towson University
Dan Olsen, Brigham Young University
Jennifer Preece, University of Maryland
Dan Siewiorek, Carnegie Mellon University
Terry Winograd, Stanford University
PCAST Statement revised
-
6 -
Ben Shneiderman 9/23/2006